Revisiting reference intervals in transfusion medicine: Innovations and challenges in current practice

Area of Science:

• Transfusion Medicine
• Clinical Chemistry
• Biostatistics

Background:

• Reference intervals (RIs) are vital for donor selection, blood component quality, and post-transfusion safety.
• The widely used CLSI EP28 non-parametric method has limitations with outliers and skewed data.
• Emerging computational methods offer potential improvements in RI estimation.

Purpose of the Study:

• Evaluate limitations of CLSI EP28 in transfusion medicine.
• Compare EM, reflimR, and refineR for refining RI estimation.
• Assess the effectiveness of computational approaches for accurate RIs.

Main Methods:

• Simulated blood donor dataset (n=500) for hemoglobin, hematocrit, and platelet count.
• Compared four RI estimation methods: CLSI EP28, Expectation-Maximization (EM), reflimR, and refineR.
• Analyzed outlier resistance and distributional accuracy of each method.

Main Results:

• CLSI EP28 and EM produced similar RIs based on empirical percentiles.
• ReflimR and refineR generated wider RIs, demonstrating improved outlier resistance.
• ReflimR and refineR showed enhanced distributional accuracy compared to CLSI EP28 and EM.

Conclusions:

• While CLSI EP28 is the standard, computational methods like EM, reflimR, and refineR offer superior accuracy and robustness.
• These advanced methods improve regulatory compliance (IVDR, FDA 510(k), ISO 15189).
• Implementing EM, reflimR, and refineR can enhance donor screening and transfusion safety.